Journal: bioRxiv

Article Title: Extensive length and homology dependent chimerism in pool-packaged AAV libraries

doi: 10.1101/2025.01.14.632594

Figure Lengend Snippet: (a) A complex doubly barcoded cloning dock with an a priori determined valid barcode pairs dictionary serves as the starting point to clone libraries of barcoded inserts varying length and class (homologous & non-homologous) within an AAV cargo. (b) Barcoded libraries were separately packaged in AAVs, and submitted for direct long-read sequencing, in parallel with corresponding insert size-selected from the plasmid digest. (c) Quantification of the fraction of discordant barcode pairs as a function of the full-length BC-to-BC average size. Each point corresponds to swap quantification from 6 separate libraries ([short, mid, long] × [homologous, non-homologous]) for both plasmid-derived (square) and AAV-derived (circle) material (n=1 replicate per library). For each datapoint, we analyzed full length BC-to-BC reads passing quality control filters and having separately valid BC1 and BC2 ( , ). Right panel corresponds to a zoomed inset showing the y axis range from 0 to 0.1. Quantifications derived from library p153 × corresponding to the pool of multiply-sized inserts in a single AAV-packaged sample are marked by an ×. Vertical error-bars correspond to 20 th to 80 th percentiles from bootstrap resampling. Horizontal error-bars to the 10th to 90th percentile of the BC-to-BC length distribution from plasmid digest inserts. Swaps are significantly higher (one-sided bootstrap FDR: ***<10 −5 ) for the homologous vs. their respective size-matched non-homologous libraries. (d) Same as (c), but with additional packaging conditions with library p153 × corresponding to AAV2 serotype and sparse co-packaging with carrier DNA. With both PHP.eB and AAV2, and for the three sizes of inserts, sparse packaging reduces swapping significantly (one-sided bootstrap FDR: *<0.005, **<0.0005, ***<10 −5 ) albeit incompletely. (e) Comparison of the fraction of discordant barcode pairs from PCR-derived libraries (x-axis) vs. direct [same data as panel (c)] (y-axis), see also . AAV samples only include the PHP.eB serotype with standard packaging conditions. Errorbars correspond to 20 th to 80 th percentiles from bootstrap resampling of reads as in panel (c).

Article Snippet: PCR-free libraries of AAV-packaged DNA were sequenced using the AAV service from Plasmidsaurus (project ID RP88L6).

Techniques: Cloning, Sequencing, Plasmid Preparation, Derivative Assay, Control, Comparison

Journal: bioRxiv

Article Title: Extensive length and homology dependent chimerism in pool-packaged AAV libraries

doi: 10.1101/2025.01.14.632594

Figure Lengend Snippet: Similar to with related modifications. (a) Table showing long-read counts retention at different filtering steps in our pipeline (schematic at right indicate filter nature), e.g., presence of 4/4 signpost sequences and (separate) exact matches to BC1 and BC2 from our dictionary. (b) Cumulative distribution of alignment scores to the signposts. Thresholds for detection (75% match) are shown as dashed lines at top. (c) Two-dimensional histogram showing the Levenshtein distance to the best and second best matches to insert indices. Decision boundary to deem insert demultiplexing as unambiguous is shown by the red dashed line. (d) Cumulative distribution of the BC-to-BC length (from middle of both barcodes) as determined by the detected signpost positions from the reads passing signpost quality control steps and with separate exact matches to BC1 and BC2. Each panel comes from a separate AAV packaging sample, and was indexed separately (Plasmidsaurus) for ONT sequencing. For p153 × samples, different inserts were demultiplexed using the internal insert index. Line type is related to insert size category (parental, short, mid-sized, long), color to insert class (black: parental, light blue: homologous, orange: non-homologous). Vertical lines indicate 1% and 99% percentiles from the plasmid libraries (same as ) used to call a read ‘full BC-to-BC length’. Note the substantial fraction of reads with shorter than expected lengths for the long insert libraries. (e) Similar to panel (d), but for total read length. Vertical lines used to consider whether the read is full ITR-to-ITR or not (p153 × -short: 500 to 750 nt, p153 × -mid: 1100 to 1400 nt, all others: 2250 to 2750 nt) for analysis of Table S3. (f) Table of counts of valid reads stratified by demultiplexed libraries (based on insert index), tallying proportion of reads with valid barcode pairs. Quantiles (1%, 10%, 90%, 99%) of the BC-to-BC length across the libraries is also shown. (g) Plasmidsaurus (order ID L9RV6B) virtual gel for confirmation of p153, detecting only a single product. (h) Agarose gel (low and high contrast) of undigested p153, showing evidence of possible lower molecular weight products (arrows). (i) PCR from both AAV template from p153 [serotype PHP.eB, standard condition, sample 5] (left) and plasmid (right) with respectively primers o949+o950 (ITR-to-ITR) and o949+o936 (BC-to-BC). We note that band intensities are probably not representative of species abundance due to possible length-bias of amplification. The three products are of the expected sizes for libraries with short, mid-sized, and long inserts respectively, consistent with their detection in the long-read data.

Article Snippet: PCR-free libraries of AAV-packaged DNA were sequenced using the AAV service from Plasmidsaurus (project ID RP88L6).

Techniques: Control, Sequencing, Plasmid Preparation, Agarose Gel Electrophoresis, Molecular Weight, Amplification

Journal: bioRxiv

Article Title: Extensive length and homology dependent chimerism in pool-packaged AAV libraries

doi: 10.1101/2025.01.14.632594

Figure Lengend Snippet: (a) Flowchart illustrating the different types of libraries considered (starting material: plasmids & AAV packaged DNA), method (PCR-based vs. direct, i.e., PCR-free, corresponding to digestion & size selection for plasmids, and annealing fpr AAVs, prior to end-repair & adapter ligation). PCR libraries were pooled prior to submission and library identity assigned by demultiplexing based on the internal index (see ). (b) Similar to , now also showing the PCR-derived libraries (open symbols). This data is represented as a comparison between direct/PCR libraries in . (c) and (d) Table showing long-read counts retention at different filtering steps in our pipeline for PCR libraries prepared from AAV packaged DNA and plasmids respectively. (e) Cumulative distribution of the BC-to-BC length (from middle of both barcodes) as determined by the detected signpost positions from the reads passing signpost quality control steps and with separate exact matches to BC1 and BC2. Line type corresponds to template type (full: plasmid, dashed: AAV), color to insert class (black: parental, light blue: homologous, orange: non-homologous). Vertical lines indicate 1% and 99% percentiles of the length distributions respectively from the plasmid digest (and used to consider a long-read in the AAV-packaged data as ‘full BC-to-BC length’). Panels correspond to different insert libraries. Parental inserts from the AAV sample is not shown as it cannot be reliably attributed to a sample (came from residual empty plasmids packaged from all libraries). (f) Similar to panel (e), but for the full read length. Grey arrows indicate ‘dimer reads’ (see Methods ) that were roughly twice the length of the expected full length reads. (g) Tapestation (D5000) of pooled PCR-prepared libraries submitted for ONT sequencing, showing predominant expected product (≈2.2 kb), and minor empty product from p146 (≈900 bp), and lack of product >4 kb corresponding to ‘dimer reads’ seen in (f). (h) Table of read counts with valid barcode pairs and valid fraction across libraries originating from plasmids or AAV packaged DNA (different rows), and direct (PCR-free) [central columns] vs. PCR-derived libraries [right columns]. The data from the direct libraries is reproduced from and for plasmid digest and AAVs respectively.

Article Snippet: PCR-free libraries of AAV-packaged DNA were sequenced using the AAV service from Plasmidsaurus (project ID RP88L6).

Techniques: Size Selection, Adapter Ligation, Derivative Assay, Comparison, Control, Plasmid Preparation, Sequencing